High-Affinity Rb Binding, p53 Inhibition, Subcellular Localization,

and Transformation by Wild-Type or Tumor-Derived Shortened

Merkel Cell Polyomavirus Large T Antigens
Sophie Borchert,a,b Manja Czech-Sioli,a Friederike Neumann,a Claudia Schmidt,a Peter Wimmer,b Thomas Dobner,b Adam Grundhoff,b
Nicole Fischera
Institute for Medical Microbiology and Virology, University Medical Center Eppendorf, Hamburg, Germanya; Heinrich Pette Institute, Leibniz Institute for Experimental
Virology, Hamburg, Germany

ABSTRACT
Interference with tumor suppressor pathways by polyomavirus-encoded tumor antigens (T-Ags) can result in transformation.

Downloaded from http://jvi.asm.org/ on June 17, 2019 by guest

Consequently, it is thought that T-Ags encoded by Merkel cell polyomavirus (MCPyV), a virus integrated in ⬃90% of all Merkel
cell carcinoma (MCC) cases, are major contributors to tumorigenesis. The MCPyV large T-Ag (LT-Ag) has preserved the key
functional domains present in all family members but has also acquired unique regions that flank the LxCxE motif. As these re-
gions may mediate unique functions, or may modulate those shared with T-Ags of other polyomaviruses, functional studies of
MCPyV T-Ags are required. Here, we have performed a comparative study of full-length or MCC-derived truncated LT-Ags with
regard to their biochemical characteristics, their ability to bind to retinoblastoma (Rb) and p53 proteins, and their transforming
potential. We provide evidence that full-length MCPyV LT-Ag may not directly bind to p53 but nevertheless can significantly
reduce p53-dependent transcription in reporter assays. Although early region expression constructs harboring either full-length
or MCC-derived truncated LT-Ag genes can transform primary baby rat kidney cells, truncated LT-Ags do not bind to p53 or
reduce p53-dependent transcription. Interestingly, shortened LT-Ags exhibit a very high binding affinity for Rb, as shown by
coimmunoprecipitation and in vitro binding studies. Additionally, we show that truncated MCPyV LT-Ag proteins are ex-
pressed at higher levels than those for the wild-type protein and are able to partially relocalize Rb to the cytoplasm, indicating
that truncated LT proteins may have gained additional features that distinguish them from the full-length protein.

IMPORTANCE
MCPyV is one of the 12 known polyomaviruses that naturally infect humans. Among these, it is of particular interest since it is
the only human polyomavirus known to be involved in tumorigenesis. MCPyV is thought to be causally linked to MCC, a rare
skin tumor. In these tumors, viral DNA is monoclonally integrated into the genome of the tumor cells in up to 90% of all MCC
cases, and the integrated MCV genomes, furthermore, harbor signature mutations in the so-called early region that selectively
abrogate viral replication while preserving cell cycle deregulating functions of the virus. This study describes comparative stud-
ies of early region T-Ag protein characteristics, their ability to bind to Rb and p53, and their transforming potential.

M erkel cell polyomavirus (MCPyV) is one of 12 human poly-

omaviruses (1, 2), and to date is the only human polyoma-
virus for which solid evidence of a causative role in tumorigenesis
exists. The virus was identified in Merkel cell carcinoma (MCC), a
rare form of skin cancer seen in elderly and immunosuppressed
patients (3). The high frequency of MCPyV detection in 60 to 90%
of all MCC cases (4–9), monoclonal integration of the viral DNA
in the tumor cells of primary tumors as well as metastases, MCC-
specific signature mutations in the viral genome, and constitutive
expression of putative viral oncogenes within the tumor cells
strongly suggest a causative role for the virus during MCC patho-
genesis (3, 9, 10).
Although most polyomaviruses do not induce tumors in their
natural host, many family members can induce transformation of
cells in vitro, with some of these viruses also being able to induce
tumors in experimental animal models. The transformation abil-
ity of polyomaviruses is intricately linked to the expression of the
early viral antigens, which include the small and large T antigens
(sT-Ags and LT-Ags, respectively). Several members of the family
additionally encode a middle T antigen (MT-Ag). MT-Ags can
likewise contribute to transformation and in some cases (e.g.,
mouse polyomavirus) even exhibit the majority of transformation
capacity (11, 12). In many polyomaviruses, however, LT-Ag ap-
pears to be the major transforming factor. LT-Ag-dependent
transformation has been extensively studied using simian virus 40
(SV40) as a model system (13). SV40 LT-Ag deregulates cell
growth by a variety of mechanisms that include binding of Hsc70,
retinoblastoma (Rb) proteins, and p53 and result in the release of
E2F-DP (E2 promoter binding-protein dimerization partner)
from hypophosphorylated Rb and inhibition of p53-mediated
transcriptional activation. At least three conserved domains con-
tribute to SV40 LT-Ag-induced transformation: the J domain and
LxCxE motif, which are located in the in the N-terminal half of the
protein and bind to Hsc70 and Rb proteins, respectively, and a p53
Received 4 October 2013 Accepted 20 December 2013
Published ahead of print 26 December 2013
Editor: M. J. Imperiale
Address correspondence to Nicole Fischer, nfischer@uke.de, or Adam Grundhoff,
adam.grundhoff@hpi.uni-hamburg.de.
Copyright © 2014, American Society for Microbiology. All Rights Reserved.
doi:10.1128/JVI.02916-13

3144 jvi.asm.org Journal of Virology p. 3144 –3160 March 2014 Volume 88 Number 6

binding domain located in the C terminus. The N terminus, in-
cluding the J domain and the LxCxE motif, can be sufficient to
transform cells in vitro (14). Similar to SV40 LT-Ag, the LT pro-
teins encoded by the related JC and BK polyomaviruses have also
been shown to induce transformation in vitro, although less effi-
ciently than those of SV40 (15, 16). The lower transformation
efficiency is a likely consequence of the reduced affinity with
which these proteins bind to the Rb family of tumor suppressors
(17).
Interestingly, all large tumor antigen (LT-Ag) sequences recov-
ered from primary MCC tumors or tumor-derived cell lines har-
bor signature mutations that lead to the expression of shortened
LT-Ag proteins. These truncation products unequivocally pre-
serve the Rb binding motif, but they lack a functional origin bind-
ing domain (OBD) and ATPase/helicase region (10, 18, 19). It is
thought that these truncations are the result of a selection pressure
to abrogate untimely firing of viral origins from integrated viral
genomes (9, 20). However, whether MCPyV LT-Ag is able to in-
terfere with p53 and whether the truncation event would abrogate
this ability are currently under discussion (21). Likewise, given
that the coding regions for sT- and LT-Ags partially overlap, it is
unclear whether retention of the LxCxE motif signifies an impor-
tant role for LT-Ag-mediated Rb inactivation during MCC patho-
genesis or, rather, reflects a requirement to preserve a functional
sT gene. In several polyomaviruses, sT-Ags can complement the
transforming and tumor-inducing functions of LT and MT (22–
24). Indeed, Shuda and colleagues demonstrated that MCPyV
sT-Ag is expressed in nearly all MCPyV-positive tumors and ex-
hibits transforming potential in rat-1 cells, most likely through a
mechanism that affects the phosphorylation status of 4E-BP1
(25). However, the fact that suppression of sT-Ag alone in sT- and
LT-Ag-positive cell lines does not fully recapitulate a pan-T
knockdown suggests a synergistic role of both T antigens during
MCC tumorigenesis (25, 26). Additionally, it was recently dem-
onstrated that sT-Ag contributes to LT-Ag expression by targeting
the cellular ubiquitin ligase SCF(Fbw7), thereby preventing pro-
teasomal degradation of LT-Ag (27).
Given the above, we have performed a side-by-side compari-
son of a consensus, full-length MCPyV LT-Ag (28) and several
truncated LT protein genes from previously characterized MCC-
derived cell lines (18). Here, we demonstrate the characterization
of the encoded LT-Ag proteins with regard to their subcellular
localization, steady-state expression, transformation potential,
and capacity to bind to Rb and p53. All analyses were performed in
comparison to the well-characterized SV40 LT-Ag protein. We
show that full-length MCPyV LT-Ag does not directly bind to p53,
but nevertheless can target the p53 pathway via an as-yet-un-
known factor, thereby significantly reducing p53-dependent tran-
scription. In contrast, tumor-derived truncated LT-Ags are un-
able to bind to p53 and do not affect p53-dependent transcription.
Our results also demonstrate that tumor-derived truncated
MCPyV LT-Ag proteins show elevated steady-state expression
levels, bind with very high affinity to Rb, and can partially relocal-
ize Rb to the cytoplasm, indicating that removal of the carboxy-
terminal region not only abrogates viral replication and lowers
growth inhibition (21, 29) but may also modulate other functions
of MCPyV LT-Ag. We also provide statistical evidence suggesting
that MCC-specific mutations positively select for the presence of
the LT-Ag LxCxE motif, independently of the need to preserve the
first exon shared with sT-Ag, and demonstrate transformation of
March 2014 Volume 88 Number 6
Functional Characterization of the MCPyV LT Antigen
primary baby rat kidney cells by early region constructs encoding
full-length and truncated MCPyV LT genes. Collectively, our data
support the hypothesis that truncated MCPyV LT-Ag plays an
important role during the pathogenesis of MCC.
MATERIALS AND METHODS
Cell lines. 293 cells (30), H1299 cells (31), and Saos-2 cells (32) were
grown as monolayer cultures in Dulbecco’s modified Eagle’s medium
(DMEM) supplemented with 10% fetal calf serum (FCS) and 5% penicil-
lin-streptomycin in a 5% CO2 atmosphere at 37°C. PFSK-1 cells, WaGa
cells, and MKL-1 cells were grown in RPMI medium supplemented with
10% FCS and 5% penicillin-streptomycin.
Plasmids and transient transfections. The MCPyV LT full-length ex-
pression construct (pCMV2b-MCPyVLTFL) was generated by PCR am-
plification using a taqExpand PCR amplification system (Roche Diagnos-
Downloaded from http://jvi.asm.org/ on June 17, 2019 by guest
tics) and genomic DNA of a synthetic MCPyV clone (MCVSyn) as the
template (28). pCMV2b-LTMCCL-3, pCMV2b-LTMCCL-11, and pCMV2b-
LTMCCL-12 were constructed by PCR amplification from genomic DNA of
MCPyV-positive MCC cell lines (MCCL-3, -11, -12) (18). pCMV2b-
LTMCC350 was generated by PCR amplification using expression plasmid
pCDNA-MCV350 (NIH AIDS Reagent Program). pCMV2b MCPyV LT
expression constructs contain an N-terminal FLAG expression tag. The
following PCR primers were used: MCPyV_XhoI_FL_R (where R indi-
cates reverse), 5=-GGGCTCGAGTTGAGAAAAAGTACCAGAATCTTG;
MCPyV EcoRV_F (where F indicates forward), 5=-CCGATATCATGGA
TTTAGTCCTAAATAGG; MCPyV_350_XhoI_R, 5=-GGGCTCGAGAT
CTGTAAACTGAGATGACGAGGC; MCCL-12_LT_XhoR, 5=-GGGCT
CGAGGAATGGAGGAGGGGTCTTCGG; SV40ori_F, 5=-CCTCGAGA
GCCTAGGCCTCCAAAAAAG; SV40_ori_R, 5=-CCTCGAGTCCGC
CCCATGGCTGAC; MCPyV_ori_F, 5=-CCTCGAGCAAGGGCGGGA
AAC; MCPyV_ori_R, 5=-CCTCGAGCTTGTCTATATGCAG; SV-YFP-
N-F-XhoI, 5=-CTCAGATCTCGAGCTATGGATAAAGTTTTAAACAGA
GAGG; SV-YFP-N-R-EcoRI, 5=-CAGAATTCTTATGTTTCAGGTTCAG
GGGG; SV-YFP-C-F-NheI, 5=-CCGCTAGCATGGATAAAGTTTTAAA
CAGAGAGG; SV-YFP-C-R-AgeI, 5=-GACCGGTGCACCTGCTCCTG-T
TTCAGGTTCAGGGGGAG; MCPyV YFP-N-F-BsrGI, 5=-CTGTACAAG
GGAGCAGGTGCAGGAGCAATGGATTTAGTCCTAAATAGG; MCPyV
YFP-N-R-NotI, 5=-GCGGCCGCTTATTGAGAAAAAGTACCAGAATC;
MCPyV YFP-C-F-NheI, 5=-CCGCTAGCATGGATTTAGTCCTAAATA
GGAAAG; MCPyV_YFP-C-R-AgeI, 5=-GACCGGTGCACCTGCTCCTT
GAGAAAAAGTACCAGAATC; MCCL-11 YFP-N-R-NotI, 5=-CGCGGC
CGCTTCATCTGGGCTGCCGGG; MCCL-11 YFP-C-R-AgeI, 5=-GACC
GGTGCACCTGCTCCTCTGGGCTGCCGGGGCG. The SV40 LT cDNA
containing plasmid pZIPTEX (33) was kindly provided by W. Deppert.
pHCMV2-SV40LTFL encoding untagged full-length SV40 LT was gener-
ated by subcloning a BamHI fragment of pZIPTEX into pHCMV (34).
Luciferase expression plasmids containing E2F binding sites (pI-H2A-
68) or mutated E2F binding sites (pI-H2A-68*) in the H2A promoter
region together with cytomegalovirus (CMV) expression plasmids for Rb
or E2F have been described before (35). Human adenovirus 5 (hAdV5)
pcDNA3.1-E1B-55K has also been described previously (36). pC53-SN3
expresses human wild-type (wt) p53 from the pCMV/neo vector (37). The
firefly luciferase reporter plasmid pRE-Luc contains five p53-binding sites
upstream of a minimal cytomegalovirus promoter and was described be-
fore (38).
pCR2.1-MCPyVori was constructed by amplification of the 97-bp
MCPyV replication origin (39) followed by ligation into the XhoI site of
pCR2.1. For transient transfection of 293 cells and H1299 cells, 25-kDa
linear polyethylenimine (Polysciences Inc., Eppelheim, Germany) was
used as the transfection reagent (40). Saos-2 cells were transfected in the
presence of FuGene (Roche).
Viral replication assays. Replication assays were performed with
PFSK-1 cells. The cells were cotransfected with 1 ␮g of pCR2.1 ori plasmid
and 1 ␮g of LT expression constructs for 48 or 96 h. Low-molecular-
weight DNA was isolated according to the Hirt protocol as recently de-
jvi.asm.org 3145
Borchert et al.
scribed (28). Briefly, DNA was digested with DpnI to distinguish repli-
cated DNA from input DNA and EcoRI for linearization. Digested DNA
was resolved on a 1% agarose gel, transferred to Hybond-N⫹ (Amersham
Pharmacia) membrane by Southern blotting, UV cross-linked, and hy-
bridized overnight at 42°C with a 32P-labeled probe (Rediprime II DNA
labeling system, random-primed labeled; GE Healthcare) corresponding
to the restriction fragment HinDIII/EcoRV (pCR2.1-MCPyVori).
Antibodies and Western blot analysis. For protein extraction, pel-
leted cells were resuspended in lysis buffer (50 mM Tris [pH 8.0], 150 mM
NaCl, 1% NP-40, 0.5% Na-deoxycholate, 5 mM EDTA, 0.1% SDS). Pro-
teins were separated by SDS-PAGE and blotted on a nitrocellulose mem-
brane. Primary antibodies specific for polyomavirus proteins used in this
study include MCPyV LT-Ag mouse monoclonal antibody (MAb)
Cm2B4 (10) (Santa Cruz) and SV40 LT mouse MAb Pab419 (41). p53
mouse MAb DO-1, polyclonal Ab FL-393, Brd4 specific rabbit polyclonal
Ab H-250, and Rb rabbit polyclonal Ab M-153 were purchased from Santa
Cruz, and the FLAG-M2 used in immunoprecipitation studies was pur-
chased from Sigma-Aldrich. Actin mouse MAb was applied in Western
blot analyses to ensure loading of equal protein amounts (Chemicon cat-
alog no. 1501). The secondary antibodies conjugated to horseradish per-
oxidase (HRP) for detection of proteins by immunoblotting were anti-
mouse IgG (GE Healthcare) and anti-rabbit IgG (Santa Cruz).
pBRK cell transformation assay and anchorage-independent
growth assay. Primary baby rat kidney (pBRK) cells were obtained from
kidneys of 7-day-old Sprague-Dawley rats as previously described (42).
Soft agar colony formation assay was performed according to published
protocols (43).
Immunofluorescence staining and confocal microscopy. For indi-
rect immunofluorescence, cells were grown on glass coverslips coated
with 0.2% gelatin (Sigma). Cells were fixed in 4% paraformaldehyde in
phosphate-buffered saline (PBS; 15 min) and permeabilized in PBS–1%
Triton X-100 for 30 min. After 1 h of blocking in Ca2⫹/Mg2⫹-free PBS
containing a buffer of 1% Triton X-100, 0.5% Tween, and 3% bovine
serum albumin (BSA; albumin fraction), coverslips were treated for 1h
with the primary antibody diluted in blocking buffer and washed in PBS–
0.1% Tween 20, followed by incubation with the corresponding second-
ary antibodies (Dianova, Hamburg, Germany). Coverslips were mounted
in Vectashield medium, and digital images were acquired with a confocal
laser-scanning microscope (Leica DM IRE2 with a Leica TCS SP2 AOBS
confocal point scanner) equipped with an oil-immersion Plan Apo ⫻63,
numerical aperture (NA) 1.4 objective. Images were cropped using Adobe
Photoshop CS5 and assembled with Microsoft PowerPoint.
Luciferase assay. For p53-dependent transcription using dual-lucif-
erase assays, subconfluent H1299 cells were transfected in 24-well plates
with a DNA/polyethyleneimine (PEI) mixture using the effector plasmids,
pc53-SN (25 ng) and pRL-TK (250 ng) (Promega). Where necessary,
empty pCMV2b vector was added such that the total amount of trans-
fected DNA was 3.5 ␮g per sample across all experiments. Total-cell ex-
tracts were prepared 36 h after transfection in 1⫻ lysis buffer, and lucif-
erase activity was assayed with 20 ␮l of extract. All samples were
normalized for transfection efficiency by measuring Renilla luciferase ac-
tivity. All experiments were performed in triplicate.
For luciferase assays measuring Rb binding and E2F activation, 3 ⫻
104 Saos-2 cells were transfected in 24-well plates using FuGene transfec-
tion reagents (Roche) with pI-H2A-68, CMV-Rb, and LT-Ag as indicated
in the legend to Fig. 6. pRL-TK was cotransfected for normalization. At 36
h posttransfection, cell extracts were prepared and luciferase activity was
determined using a dual-luciferase assay (Promega) according to the
manufacturer’s instructions.
Coimmunoprecipitation (co-IP) studies. Total-cell extracts were
prepared by using lysis buffer (10 mM HEPES [pH 7.8], 10 mM KCl, 2
mM MgCl2, 0.1 mM EDTA, 1% Nonidet P-40) supplemented with a
protease inhibitor mixture (Roche). After 30 min on ice and cell disrup-
tion, 2 volumes of TN buffer (200 mM NaCl, 20 mM Tris supplemented
with protease inhibitors) were added, and cell lysate was cleared by cen-
3146 jvi.asm.org
trifugation (4°C, 14,000 rpm, 30 min). After normalizing for protein con-
centration, whole-cell extracts were subjected to immunoprecipitation.
Supernatant was precleared by adding 35 ␮l protein A/G Sepharose (Santa
Cruz) for 30 min at 4°C. A 10-␮l antibody solution was added to the lysate,
and the mixture was rotated at 4°C O/N. Thirty-five microliters of protein
A/G Sepharose was added for 1 h at 4°C with rotation. Beads were washed
5⫻ (TN buffer), and FLAG proteins were eluted from the beads by com-
petition with the FLAG peptide (100 ␮g/ml) or by adding SDS loading
buffer and with subsequent boiling of the samples.
FACS-FRET. Fluorescence-activated cell sorter–fluorescence reso-
nance energy transfer (FACS-FRET) was performed as recently described
(44) using the Clontech vectors pEYFP-C1/N1 and pECFP-C1/N1.
Merkel cell polyomavirus sequences were N- and C-terminally tagged
using the pEYFP-N1 vector and the restriction sites NheI/AgeI and BsrGI/
NotI, respectively. Tagged SV40 LT-Ag was generated using the
pEYFP-C1 vector with restriction sites NheI/AgeI for N-terminal tagging
Downloaded from http://jvi.asm.org/ on June 17, 2019 by guest
and the vector pECFP-C1 with XhoI/EcoRI sites for C-terminal tagging.
The positive control for FACS-FRET (cyan fluorescent protein [CFP]-
fused yellow fluorescent protein [YFP]) was kindly provided by M. Schin-
dler (44). pEYFP p53 and pECFP p53 plasmids were provided by W.
Ching, Heinrich Pette Institute, Hamburg.
Quantitative RT-PCR. Total RNA from three independent experi-
ments was isolated from PFSK-1 or U2OS cells at 48 h posttransfection
using an RNeasy Mini Kit and column DNase I digestion. One hundred
nanograms of total RNA was used in a one-step reverse transcription-PCR
(RT-PCR; Thermo Scientific) using gene-specific primers in a 10-␮l vol-
ume of reaction mixture. Primer sequences for p21, Hdm2, GADD45A,
cdc2, cyclA2, glyceraldehyde-3-phosphate dehydrogenase (GAPDH), and
RLP13 have been published previously (29, 45, 46).
Cycling was performed using a Rotor-Gene Q-plex (Qiagen) at the
following conditions: 5 min at 95°C, 40 cycles of 15 s at 95°C, 20 s at 60°C,
and 15 s at 72°C, followed by melting curve analysis. The data were ana-
lyzed using Rotor-Gene software, and mRNA of each gene was normal-
ized to two housekeeping mRNA levels, those of GAPDH and RLP13.
Cell fractionation. To fractionate cells into cytoplasm and nuclei, cells
were resuspended in 10 mM HEPES [pH 7.9], 1.5 ml MgCl2, 10 mM KCl,
and 0.5 mM dithiothreitol (DTT) and incubated on ice for 5 min. Cells
were disrupted using a Dounce homogenizer. The cytoplasmic fraction
was separated from nuclei and other components by centrifugation at
228 ⫻ g at 4°C for 5 min. To isolate nuclei, the nuclear pellet was resus-
pended in 0.25 M sucrose–10 mM MgCl2 and centrifuged for 10 min at
2,800 ⫻ g at 4°C through a sucrose cushion containing 0.88 M sucrose and
0.5 mM MgCl2. The method used to fractionate cells into the cytoplasm,
nucleoplasm, nuclear membrane, chromatin, and extracellular matrix
was published earlier (47).
Bacterial expression and purification of N-terminal LT proteins and
the Rb pocket domain. Coding sequences for truncated SV40 LT7-117 and
MCPyV LT1-244 were PCR amplified using pZIPTEX (33) or cDNA from
the MCPyV-positive MCC cell line 12 (18) as the template, and PCR
fragments were sequenced and subcloned into pRSETA (Invitrogen). N-
terminal His-tagged proteins were expressed in BL21 Star in LB medium
supplemented with 1% glucose at 37°C for 4 h using 1 mM isopropyl-␤-
D-thiogalactopyranoside (IPTG; for SV40 LT) or at 16°C for 18 h using 0.5
mM IPTG (for MCPyV LT). Ni-nitrilotriacetic acid (Ni-NTA) purifica-
tion was performed with 50 mM Tris [pH 7.5], 150 mM NaCl (750 mM
NaCl for MCPyV LT), 5% glycerol, and 1 mM DTT (5 mM ␤-mercapto-
ethanol for MCPyV LT). Proteins were eluted using an imidazole gradient
of 20 to 500 mM imidazole. Size exclusion chromatography was executed
under the same buffer conditions with an Äkta Prime Superdex 200. Frac-
tions containing monomeric LT protein were pooled, dialyzed, and con-
centrated. Rb sequences encoding to the pocket domain (amino acids
[aa]382 to 771) were subcloned into the glutathione S-transferase (GST)
expression plasmid pGEX4T3. Bacterial expression was performed with
BL21 Star grown in LB medium supplemented with 1% glucose, protein
Journal of Virology

expression was induced with 0.5 mM IPTG, and cells were grown at 16°C
for 18 h. The purification strategy has been described before (48).
MST. The microscale thermophoresis (MST) background has been
reviewed recently (49). The dissociation constant of bacterially expressed
N-terminally truncated LT proteins binding to the purified Rb pocket
domain was measured using a Monolith NT115 (NanoTemper). Purified
Rb pocket domain fragments were Alexa 647 labeled by following the
manufacturer’s protocol. A solution of monomeric LT proteins was seri-
ally diluted as indicated in Results. The labeled Rb pocket domain was
incubated with the LT protein dilutions at room temperature in the dark
for 30 min. Measurements were performed with 50 mM Tris [pH 7.4],
0.05% Tween [150 mM], and 200 mM or 250 mM NaCl at 50% red
light-emitting diode (LED) power and 60% infrared (IR) laser power
using hydrophilic capillaries. Measurements were also carried out at 20%
and 40% IR laser power. Three independent measurements at the condi-
tions described in Results were executed. Data analysis was performed
using NanoTemper analysis software. Curves were fitted using nonlinear
regression Hill slope (GraphPad Prism).
Statistical analysis of MCC-derived MCPyV LT-Ag genes. In addi-
tion to three truncated LT-Ag genes analyzed in one of our earlier studies,
we extracted all annotated MCPyV LT-Ag coding sequences deposited in
the NCBI database as of 15 January 2013 (n ⫽ 148). After multiple nucle-
otide sequence alignments, we identified all MCC-derived entries that
contain a truncating mutation and eliminated all duplicate entries, leaving
40 unique entries for further analysis (detailed information is available
upon request). For amino acids encoded by the second exon (the first
exon must be considered fixed, since it is shared with sT-Ag), we next
calculated P values for the hypothesis that this position is subject to pos-
itive or negative selection during MCC pathogenesis. For negative selec-
tion of amino acids located between the first observed truncation site and
the last amino acid encoded by the second exon, the probability of the
corresponding null hypothesis is given by the accumulated binomial
probability that exactly the observed number of or more truncation events
occur before the given position. Conversely, for positive selection of
amino acids between the first amino acid encoded by the second exon and
the last observed truncation site, the probability of the null hypothesis
corresponds to the accumulated probability that exactly the observed
number of or more truncation events occur after the position in question.
The amino acid positions at which the significance of the calculated P
values was maximal were then considered to be most likely to signify the
position at which positive or negative selection, respectively, occurs.
RESULTS
MCC-derived truncated LT-Ags show highly significant selec-
tion for the presence of the LxCxE motif. All known MCC-de-
rived truncated LT-Ag (tLT-Ag) harbor mutations that lead to the
removal of the C-terminal helicase/ATPase domain, presumably
due to a selection pressure to abrogate untimely firing of viral
origins (9, 20). It is also notable that the truncating mutations
unequivocally occur downstream of the LxCxE motif, an observa-
tion which would seem to suggest an important role for tLT-Ag-
mediated Rb inactivation during MCC pathogenesis. Yet, given
that LT- and sT-Ag share their first exon, one could also argue that
the above may predominantly reflect a requirement to preserve a
functional sT-Ag gene.
Indeed, only 60 nucleotides (nt) separate the LxCxE coding
sequences from the 5= end of the second exon in the SV40 LT-Ag
gene. However, the MCPyV LT-Ag gene additionally encodes a
unique 118-aa-long serine-rich region of unknown function (Fig.
1A, U1 region) and thus offers significantly more space for muta-
tions that may remove the LxCxE motif without affecting sT-Ag.
In order to test whether the absence of such mutations in MCC-
derived tLT-Ags is statistically significant, we extracted from the
NCBI database all unique tLT-Ag entries (n ⫽ 40) and identified
March 2014 Volume 88 Number 6
Functional Characterization of the MCPyV LT Antigen
the position of the last preserved amino acid, i.e., the last amino
acid position preceding stop codon or frameshift mutations. As
shown in Fig. 1A, most mutations (n ⫽ 31; 78%) result in trunca-
tion between aa 243 and 329 and remove the origin binding do-
main (OBD), whereas only 7 mutations (18%) leave the OBD
intact. Furthermore, the 31 proximal mutations map upstream of
the 57-kDa splice donor, such that the spliced transcripts would
likewise generate a truncated LT-Ag. In 9 cases, the mutations are
located within the 57-kDa intron, and these isolates should thus be
able to produce a full-length 57-kDa T-Ag.
For each amino acid position encoded by the second exon of
the full-length LT-Ag gene, we next calculated the P value for the
hypothesis that this position is subject to either positive or nega-
tive selection in MCC (see Materials and Methods for details of the
statistical analysis). The P values for negative selection of the
Downloaded from http://jvi.asm.org/ on June 17, 2019 by guest
LT-Ag carboxy terminus were highly significant, with the lowest P
value observed for truncation after amino acid position 329 (P
value, 1.76e-18) (Fig. 1A). In addition, our analysis strongly sug-
gests positive selection for the first 165 aa encoded by the second
exon, which includes the LxCxE motif (P value, 7.83e-10). In con-
trast, we found no evidence to suggest positive selection of the
recently identified (50) functional nuclear localization signal
(NLS) sequence located between aa 277 to 280 (P value, 0.13) or
the sequences that are homologous to the NLS of SV40 LT-Ag (aa
299 to 307; P value, 0.60). Hence, the presence of these signals is
not a general requirement for tLT-Ag functions during MCC
pathogenesis.
MCC-derived truncated LT-Ags differ from full-length LT-
Ags with regard to subcellular localization and steady-state ex-
pression levels. Given that a significant number of tLT-Ags (16 of
40; 38%) lack previously identified NLSs, we sought to investigate
the expression and subcellular localization of representative
LT-Ag truncation products. We therefore generated eukaryotic
expression constructs for the full-length MCPyV LT antigen de-
rived from a consensus MCPyV genome (28) as well as three trun-
cated tLT-Ag genes isolated from MCC-derived cell lines (MCCL-
12, -11, and -3) (18) and a shortened LT-Ag from the primary
tumor MCC350 (3). As indicated in Fig. 1B, MCCL-11 is trun-
cated 27 aa downstream of the LxCxE motif and represents the
shortest LT-Ag isolate identified to date, whereas MCC350,
MCCL-12, and MCCL-3 terminate 14, 32, and 49 aa further
downstream, respectively. Of the four truncated LT-Ags, only
MCCL-3 retains the functional NLS identified by Nakamura et al.
(50). To ensure functionality of the N-terminally tagged full-
length proteins, we performed viral DNA replication assays in the
presence of an additional plasmid containing the minimal
MCPyV origin of replication (39). As shown in Fig. 2, MCPyV-LT
N-terminally fused to a FLAG tag or CFP is highly expressed and
able to support DNA replication. This observation is in line with
other studies reporting functionality of N-terminally tagged
MCPyV LT proteins in replication assays, coimmunoprecipita-
tion studies, and helicase assays (21, 29, 39). We used confocal
microscopy in combination with the monoclonal antibody
Cm2B4 (which recognizes aa 116 to 129 of LT-Ag) (9) to detect
MCPyV LT in transfected H1299 cells. As shown in Fig. 3, full-
length MCPyV LT-Ag (and SV40 LT-Ag, which was employed as a
positive control) localized strictly to the nucleus. In contrast, all
three MCC-derived LT-Ags that lack the NLS motif were distrib-
uted diffusely in the nucleus and the cytoplasm (Fig. 3). Interest-
ingly, a substantial fraction of the MCCL-3-derived tLT-Ag (tLT-
jvi.asm.org 3147
Borchert et al.

Downloaded from http://jvi.asm.org/ on June 17, 2019 by guest

FIG 1 Preservation of the LxCxE motif in MCC-derived LT-Ag proteins is highly significant. (A) Frequency plot showing the position of the last preserved amino
acid (i.e., the last position preceding stop codon or frameshift mutations) across 42 MCC-derived truncated MCPyV LT-Ag genes. (B) Amino acid alignment of
SV40 LT-Ag (GI: 297591903) and MCPyV LT-Ag (GI: 372100550) using LALIGN. Identical, conserved, and semiconserved amino acids are indicated by
asterisks, colons, and periods, respectively. The functional DnaJ, OBD, zinc finger, and helicase/ATPase domains, as well as the Rb binding LxCxE motif, are
shown as light gray boxes in panels A and B. Regions that undergo positive or negative selection according to our statistical analysis are marked. In panel A, dark
gray boxes labeled U1 and U2 symbolize regions that are unique to MCPyV and not present in SV40. The position of a recently identified functional NLS in
MCPyV LT-Ag (50) and the partially conserved NLS of SV40 are labeled NLS a and b, respectively. Underlined amino acids in panel B highlight the bipartite
regions that mediate p53 binding of SV40 LT-Ag, and individual residues that make direct contact with p53 are additionally marked by arrows (72).

3148 jvi.asm.org Journal of Virology

 Functional Characterization of the MCPyV LT Antigen

FIG 2 Viral replication assays using pCR2.1-ori plasmids and LT expression plasmids. (A) Southern blot of HIRT DNA extracts from PFSK-1 cells transfected
with 1 ␮g pCR2.1 MCPyV-ori and N-terminally or C-terminally tagged MCPyV LT expression constructs. (B) Western blot of PFSK-1 cells expressing MCPyV
LT-Ag 48 h posttransfection. Mock, mock transfection; d, days.

Downloaded from http://jvi.asm.org/ on June 17, 2019 by guest

AgMCCL-3) likewise localized to the cytoplasm, even though this
isolate retains an NLS and shows increased nuclear accumulation
compared to the other tLT-Ags (Fig. 3). All tested proteins showed
marked exclusion of nucleolar staining. To further corroborate
these findings, we performed cell fractionation experiments with
Merkel cell cancer cell lines as well as PFSK-1 cells transiently
overexpressing MCPyV LT full-length and truncated protein (Fig.
4). In accord with our immunofluorescence data, the predomi-
nant fraction of MCC-derived truncated LT proteins can be de-
tected in the cytoplasm. In WaGa cells, the truncated protein lo-
calizes almost exclusively to the cytoplasmic fraction, whereas in
MKL-1 cells, the protein is present in the nuclear and cytoplasmic
fractions (Fig. 4A). Similar results were obtained for cells with an
ectopically expressed MCCL-11-derived tLT-Ag (Fig. 4B, upper
panel). In contrast, full-length MCPyV LT-Ag was detected exclu-
sively in the chromatin-associated nuclear fraction and the insol-
uble matrix (Fig. 4B, lower panel). We conclude from the above
data that strict nuclear localization is not a general feature of
MCC-derived shortened LT-Ags.
In our localization and Western blot studies, we repeatedly
observed higher steady-state expression levels of truncated
MCPyV LT-Ag than of the full-length protein (Fig. 3 and 4B and
data not shown). Given these differences, the amounts of trans-
fected expression constructs were titrated in subsequent experi-
ments that required equal protein expression levels (Fig. 4) (dis-
cussed below). The reasons for the different expression efficiencies
are currently unclear. Our constructs harbor the entire early re-
gion of MCPyV and thus should be principally able to express sT.
Since the sT antigen was recently shown to positively influence LT
stability (27), it would seem possible that the observed differences
are due to preferential stabilization of truncated LT protein. How-
ever, we find this unlikely as we did not observe significant differ-
ences between the stability of ectopically expressed full-length and
that of truncated LT proteins in half-life experiments (data not
shown). We found the C-terminal coding region of the LT open
reading frame (ORF) to be substantially enriched in rare codons
(data not shown) and therefore suspect that translation efficiency
may be a contributing factor. However, as we did not measure
transcript abundance, it is also possible that there are differences FIG 3 Subcellular localization of full-length SV40 LTFL, full-length MCPyV FL,
on the transcriptional level. and truncated LT proteins (MCPyV tLT) overexpressed in H1299 cells. Confocal
microscopy shows LT protein localization using Pab419 MAb and Cm2B4 MAb,
Full-length MCPyV LT-Ag coprecipitates with p53 and in- fluorescein isothiocyanate (FITC)-conjugated secondary antibodies. F-actin was
hibits p53-dependent transcription. The ability of SV40 LT-Ag stained with Cy5 phalloidin. SV40 LTFL localization stained with Pab419 was used
to transform cells correlates with its ability to interact with the as a control.

March 2014 Volume 88 Number 6 jvi.asm.org 3149

Borchert et al.
FIG 4 Cell fractionation of Merkel cell cancer (MCC) cell lines and PFSK-1
cells overexpressing MCPyV LTFL or MCC-derived truncated MCPyV tLT.
(A) MCPyV-positive MCC cell lines MKL-1 and WaGa and MCPyV-negative
MCC cell line UISO were subjected to cell fractionation. Nuclear (N) and
cytoplasmic (C) extracts were analyzed by Western blotting, applying MCPyV
LT-specific Ab Cm2B4 and, as a control, a specific antibody recognizing the
nuclear protein Brd4. (B) PFSK-1 cells transiently overexpressing full-length
MCPyV LTFL and truncated MCPyV tLTMCCL-11 were analyzed by cell frac-
tionation. Cell fractions corresponding to cytoplasm (C), nuclear matrix
(NM), nucleoplasm (NP), chromatin (Chr), and insoluble matrix (M) were
analyzed by Western blotting. trunc, truncated.
tumor suppressor proteins Rb and p53 (14). As shown in Fig. 1B,
the bipartite binding region of SV40 LT-Ag as well as individual
amino acids that make contact with p53 are not well conserved in
MCPyV. In order to directly investigate the ability of full-length
MCPyV LT-Ag to bind to p53 or p53-containing complexes, we
performed coimmunoprecipitation analyses of ectopically ex-
pressed p53 and MCPyV LT-Ag in 293 (data not shown) and
H1299 (Fig. 5) cells. As shown in Fig. 5A and C, we observed that
an antibody against p53 can indeed coprecipitate full-length
MCPyV LT-Ag but not an MCC-specific truncated LT protein.
The amount of coprecipitated full-length MCPyV LT protein, how-
ever, was considerably lower than that of SV40 LT-Ag (Fig. 5B).
To examine the consequences of LT-Ag binding on p53-medi-
ated transactivation, p53-negative H1299 cells were cotransfected
3150 jvi.asm.org
Downloaded from http://jvi.asm.org/ on June 17, 2019 by guest
FIG 5 MCPyV LTFL coprecipitates p53 in transiently transfected H1299 cells,
while truncated MCPyV tLTs fail to interact with p53. Cells were cotransfected
with p53 and different LT-Ag constructs encoding full-length MCPyV LTFL
(A), full-length SV40 LTFL (B), or truncated MCPyV tLTMCCL-11 (C). Upper
blots represent immunoprecipitation of p53 using rabbit polyclonal ␣-p53
(FL-393) coupled to Sepharose. Lower blots represent coprecipitated proteins.
Ten percent of the input and 50% of the co-IP complex are shown in the each
lane.
with expression constructs for wt p53 (pCN-53), LT-Ag, and the
p53-dependent luciferase reporter construct pRE-luc. The re-
porter contains five artificial p53 binding sites, such that luciferase
activity directly correlates with p53-dependent transcription acti-
vation. The adenoviral protein E1B55K-wt (51, 52) and SV40
LT-Ag were used as positive controls. Given the differences in
expression efficiency as described above, in order to allow direct
cross-comparison between the various LT-Ags we titrated the
amount of transfected full-length and truncated MCPyV LT-Ag
expression constructs such that at the highest concentration, the
protein levels were approximately equal and comparable to those
seen for SV40 LT-Ag (Fig. 6B). As shown in Fig. 6A, SV40 LT-Ag
reduced p53-dependent transcription by 72%. Full-length
MCPyV LT-Ag likewise inhibited p53-dependent transcription in
a dose-dependent manner, although to a lesser degree (60% at the
highest expression level). In accord with the results obtained in the
coimmunoprecipitation experiments, the truncated MCPyV LT
Journal of Virology
 Functional Characterization of the MCPyV LT Antigen

Downloaded from http://jvi.asm.org/ on June 17, 2019 by guest

FIG 6 MCC-derived truncated MCPyV tLT proteins do not repress p53-dependent transcription in transiently transfected H1299 cells. Subconfluent H1299
cells were transfected with pRL-TK, pRE-Luc, pc53-SN3, and LT-Ag expression constructs (100 ng pZIPTEX-SV40LTFL, 3 ␮g pCMV2b-MCPyVLTFL, 200 ng
pCMV2b-MCPyVLTMCCL-12, 200 ng pCMV2b-MCPyVLTMCCL-11, 200 ng pCMV2b-MCPyVLTMCCL-3, 200 ng pCMV2b-MCPyVLTMCV350). DNA amounts
were adjusted with empty vector such that the total amount of transfected DNA was equal in each sample. The cells were harvested after 48 h and analyzed for
firefly luciferase activity and Renilla luciferase activity. (A) Normalized luciferase activity relative to that of the positive control (cells transfected with p53 and
reporter constructs) is shown. Adenovirus E1B-55k-wt protein efficiently repressing p53-dependent transcription was used as an internal control. P values using
an unpaired t test are indicated. (B) Western blot of cells analyzed in panel A with loading of equal amounts of LT protein expressed under these conditions.
Unspecific background binding of Cm2B4 Ab staining is indicated with an asterisk.

March 2014 Volume 88 Number 6 jvi.asm.org 3151

Borchert et al.

TABLE 1 FACS-FRET results teins in vivo (Table 1). In contrast, we did not measure any
Construct FRET signal (%)b FRET signal for N- or C-terminally tagged MCPyV LT fusion
CFP/YFP proteins (full-length or truncated) (Table 1 and data not shown)
YFP CFP distancea Expt A Expt B Expt C or for any of the various p53 fusions, indicating that MCPyV
SV40 LT N p53 N 1 0.4 1.5 1.4 LT-Ag most likely does not directly bind to p53.
SV40 LT N p53 C 1 0 0 0 MCPyV FL and truncated LT-Ag bind to Rb and release Rb-
SV40 LT C p53 N 2 24.4 20.7 29.3
dependent repression of E2F-regulated transcription. The abil-
SV40 LT C p53 C 3 76.4 80.1 71.1
ity of tumor-derived truncated full-length and LT proteins of
p53 N SV40 LT N 1 0.7 0.9 0.2
p53 C SV40 LT N 1 1.6 1.9 0.4 MCPyV to bind to Rb through the highly conserved LxCxE motif
p53 N SV40 LT C 2 41.1 40.2 44.5 has previously been confirmed by coimmunoprecipitation exper-
p53 C SV40 LT C 3 47.3 32.1 36.2 iments (9). Here, we compared coprecipitation efficiencies of Rb
MCPyV LT N p53 N 0 0 0 and SV40 LT-Ag as well as full-length and truncated MCPyV LT
MCPyV LT N p53 C 0 0 0 proteins. As in our previous p53 binding studies, we used a larger
MCPyV LT C p53 N 0 0 0 amount of the full-length MCPyV expression construct to obtain
MCPyV LT C p53 N 0 0 0 expression levels comparable to those of the truncated proteins.

Downloaded from http://jvi.asm.org/ on June 17, 2019 by guest

p53 N MCPyV LT N 0 0 0 As shown in Fig. 7A and B, the full-length MCPyV LT protein
p53 C MCPyV LT N 0 0 0
bound weakly to Rb in transiently transfected H1299 cells com-
p53 N MCPyV LT C 0 0 0
pared to results with SV40 LT-Ag. Interestingly, as shown in Fig. 7,
p53 C MCPyV LT C 0 0 0
MCPyV tLT N p53 N 0 0 0
a considerably more substantial fraction of the truncated
MCPyV tLT N p53 C 0 0 0 MCCL-11 protein was coprecipitated by the Rb antibody (com-
MCPyV tLT C p53 N 0 0 0 pare the amount of coprecipitated LT-Ag to the input in the lower
MCPyV tLT C p53 C 0 0 0 panels of Fig. 7A and C). Likewise, experiments using a FLAG
p53 N MCPyV tLT N 0 0 0 antibody to pull down LT-Ag revealed that all MCC-derived
p53 N MCPyV tLT C 0 0 0 LT-Ag proteins coprecipitated substantially larger amounts of Rb
p53 C MCPyV tLT N 0 0 0 protein than full-length protein (Fig. 7D to F). We obtained sim-
p53C MCPyV tLT C 0 0 0 ilar results with PFSK-1 cells (data not shown), suggesting that
a
Calculated distance of FRET partners using PyMOL. 1, distance of 7 to 14 nm, more efficient binding to Rb is a general feature of truncated LT
nonparallel; 2, distance of 8 to 14 nm, parallel; 3, distance of 4 to 7 nm, parallel.
b
proteins.
Three independent measurements (shown here as experiments A, B, and C) were
taken. We next investigated functional consequences of the interac-
tion between MCPyV LT-Ag proteins and Rb by measuring the
release of Rb-dependent transcriptional repression (Fig. 8). For
this purpose, we performed transcriptional derepression assays
proteins did not significantly influence transactivation of the re- using Rb-negative Saos-2 cells transfected with a luciferase re-
porter construct by p53. porter construct containing the E2F-dependent H2A promoter
Full-length MCPyV LT-Ag does not directly bind to p53. (35), expression constructs for Rb and E2F, and either an empty
LT-Ag proteins encoded by SV40, BK virus, and JC virus have vector control or expression constructs for LT proteins. SV40 LT-
been shown to directly bind to p53 (15, 53–58), whereas mouse Ag, for which release of Rb from E2F has been described before
polyomavirus LT-Ag does not bind p53 or inhibit p53-dependent (14, 60), was used as a positive control. As shown in Fig. 8, coex-
transcription (59). To investigate whether the observed coprecipi- pression of Rb and E2F in Saos-2 cells resulted in repression of
tation of MCPyV LT antigen with p53 is due to direct or indirect luciferase expression to approximately 30% of the levels seen after
binding, we performed FACS-FRET experiments with living cells, transfection of the E2F construct alone. Simultaneous expression
as recently described (44), using different cell lines that ectopically of SV40 LT-Ag resulted in a partial rescue of Rb-mediated tran-
express MCPyV LT or SV40 LT proteins and p53 as YFP or CFP scriptional repression to 50% of baseline levels. Coexpression of
fusion proteins. FRET is based upon the transfer of energy from an full-length MCPyV LT, as well as truncated MCPyV LT proteins
excited donor fluorophore to a close-by acceptor fluorophore. As (comparable protein expression levels were applied), likewise re-
this transfer occurs only when the distance between donor and lieved repression by Rb, restoring luciferase expression to 55 to
acceptor is below 8 nm, indirect interactions are much less likely 65% of that of the E2F-only control. Although derepression was
to yield a positive signal than are direct interactions. All proteins slightly more efficient for the truncated LT proteins, the differ-
were N- and C-terminally tagged in all possible combinations in ences between the individual MCPyV LT proteins were not statis-
order to minimize the risk of producing false-negative results that tically significant (P value, 0.0647; 95% confidence interval) in this
may be the consequence of tag-induced structural alterations. assay.
Confocal microscopy was performed for all p53 and LT-Ag To further follow up on these results, we performed real-time
constructs to confirm expression and proper subcellular local- RT-PCR experiments with PFSK-1 cells transiently transfected
ization patterns (data not shown). Although N-terminally with the MCPyV LT expression constructs indicated in Fig. 9.
fused SV40 LT-Ags proved nonfunctional (potentially due to Similar to previous experiments, transfected DNA amounts were
structural disturbances of the fusion product), cotransfection adjusted to ensure equal LT protein expression levels (Fig. 9B). At
of all C-terminally fused SV40 LT-Ags and p53 resulted in 48 h posttransfection, quantitative RT-PCR was performed to
highly significant FACS-FRET signals, with an average of 45% quantify expression levels of the cdc2 and cyclinA2 E2F target
for different cell lines (293, H1299, and PFSK-1 cells), thus genes. Quantification of GAPDH and RPL13 levels was used for
readily confirming the direct interaction between the two pro- normalization purposes. As shown in Fig. 9A, we find a significant

3152 jvi.asm.org Journal of Virology

 Functional Characterization of the MCPyV LT Antigen

FIG 7 Full-length MCPyV LTFL weakly binds Rb in transiently transfected H1299 cells while truncated MCPyV tLT strongly interacts with Rb. Cells were Downloaded from http://jvi.asm.org/ on June 17, 2019 by guest
cotransfected with Rb and LT expression constructs. To ensure equal levels of expression, 3 ␮g of the full-length LT and 0.5 ␮g of truncated LT expression
constructs were used. Empty vector was used to the total amount of transfected DNA such that it was equal in each sample. Cells were harvested after 72 h, and
total-cell extracts were prepared. Coimmunoprecipitation of MCPyV LTFL (A), SV40 LTFL (B), and MCPyV tLT (C) using polyclonal Rb serum for precipitation
of interacting proteins. (D) Full-length MCPyV LT FL or truncated MCPyV tLT proteins (E and F) were precipitated with an anti-FLAG antibody; coprecipitating
proteins were eluted with FLAG peptide (peptide eluate). In addition, beads were boiled in sample buffer (beads eluate). Cell lysate from H1299 cells transfected
with a vector control (Mock control) were used as a negative control in co-IP experiments. Beads were boiled in sample buffer to control for unspecific binding.

increase in cdc2 and cyclinA2 expression when truncated MCPyV
LT protein or SV40 LT protein was expressed. In contrast, MCPyV
LT full-length protein expression did not induce a significant in-
crease of cdc2 and cyclinA2 expression levels. These data suggest
that, at least under the conditions used here, truncated LT is more
potent than the full-length protein in its ability to induce tran-
scription of authentic E2F target genes.
Truncated MCPyV LT-Ag display high binding affinity to the
Rb pocket domain. To follow up on our coimmunoprecipitation
results, where we consistently observed robust binding of Rb to
truncated MCPyV-LT proteins, we sought to further investigate
parameters of the interaction by use of microscale thermophoresis
(MST), a novel method that allows the quantitative measurement
of binding affinities between proteins in solution (61). For this

March 2014 Volume 88 Number 6 jvi.asm.org 3153

Borchert et al.
FIG 8 Release of Rb-dependent repression by tumor-derived truncated
MCPyV tLT. (A) Reporter construct pI-H2A-68 was cotransfected with an Rb
expression plasmid in the presence of E2F and the indicated LT proteins into
Saos-2 cells. Luciferase activity (rel luc activity; firefly and Renilla luciferase to
normalize for transfection differences) was determined 48 h after transfection.
E2F-induced transcription activity was set to 100, and release of repression by
LT-Ag is indicated. (B) Equal levels of LT-Ag expression were confirmed by
Western blotting.
purpose, we bacterially expressed a His-tagged MCPyV LT-Ag
fragment corresponding to the MCCL-11 isolate (MCPyV
LT1-244) as well as a His-tagged amino-terminal fragment of SV40
LT-Ag spanning amino acids 7 to 117 (a region that had been
previously used in cocrystallization studies of SV40 LT-Ag and the
Rb pocket domain [62]) and purified both proteins by Ni-NTA
and size exclusion chromatography. The pocket domain of Rb was
expressed as a GST fusion protein (GST-Rb382-771) and likewise
purified by affinity chromatography, followed by thrombin cleav-
age and ion-exchange chromatography. The purified Rb protein
was fluorescently labeled and kept at a constant concentration of
33 nM, whereas the MCPyV and SV40 LT proteins were titrated
over a range of 0.13 to 531 and 0.82 to 6725 nM, respectively. As
shown in Fig. 10, at a salt concentration of 150 mM NaCl, we
observed an average binding affinity of 1950 nM between SV40 LT
and the Rb pocket domain. Interestingly, at the same salt concen-
tration, MCPyV tLT-Ag exhibited a significantly higher binding
affinity of 51.3 nM. At an elevated salt concentration of 250 mM,
MCPyV still bound to Rb with an average affinity of 121 nM
3154 jvi.asm.org
(Table 2), whereas no binding could be observed for SV40 LT-Ag
under these conditions.
Rb mislocalizes to the cytoplasm in cells expressing trun-
cated MCPyV LT protein. Considering the high-affinity between
MCPyV tLT-Ag and Rb and the fact that a substantial fraction of
NLS-containing and NLS-deficient MCPyV LT proteins localize
to the cytoplasm, we next investigated the subcellular localization
of Rb in LT-Ag-expressing cells. For this purpose, H1299 cells
were transiently transfected with Rb and SV40 or full-length or
truncated MCPyV LT-Ag expression constructs. At 48 h post-
transfection, cells were analyzed by double staining and confocal
laser scanning microscopy. As shown in Fig. 11, Rb exhibited the
expected nuclear localization in cells that express full-length SV40
or MCPyV LT protein. In contrast, upon expression of tumor-
derived truncated MCPyV LT-Ag, a significant portion of Rb
Downloaded from http://jvi.asm.org/ on June 17, 2019 by guest
shows redistribution to the cytoplasm, where it colocalizes with
shortened LT-Ag. Endogenous Rb protein shows a very similar
behavior in cells that ectopically express truncated MCPyV LT
protein (data not shown).
MCC-derived T-Ag region induces cell transformation. Ex-
pression of SV40 LT-Ags can transform primary rodent fibro-
blasts, resulting in focus formation and multilayered cell growth
of the transformed cells, whereas untransformed cells undergo
senescence and cell death after a few weeks. Human polyomavirus
BK- or JC-derived LT-Ag also transforms primary rodent fibro-
blasts, albeit with lower efficiency (63). Shuda and colleagues
demonstrated that MCPyV sT-Ag, but not MCPyV LT-Ag, in-
duces focus formation in low-passage-number immortalized
rat-1 fibroblasts (25). Recent publications using already immor-
talized rodent cells transfected with LT-Ag-expressing plasmids
showed that truncated MCPyV LT proteins support cell growth of
these cells under low-serum conditions and in anchorage-inde-
pendent growth assays (21, 29). However, transformation of pri-
mary rodent cells by MCPyV T-Ags has not been described to
date. We compared the transforming potential of SV40 LT-Ag and
full-length or shortened MCPyV T-Ag proteins expressed in the
context of the complete MCPyV early region (these constructs are
thus principally able to express sT as well as LT). Figure 12A shows
a representative result of quantitative focus formation assays in
primary baby rat kidney (pBRK) cells. The transiently transfected
SV40 and MCPyV T proteins induced substantial focus formation
in pBRK cells, leading to the outgrowth of multilayered aggregates
of pBRK cells within 28 days. Interestingly, the constructs encod-
ing tumor-derived truncated LT-Ag genes yielded similar or
higher transforming potential than that of the wt gene, although
they were unable to interfere with p53-dependent transcription,
with the exception of one isolate (MCCL-12). Individual colonies
could be isolated from all transformed cell cultures (including
MCCL-12) and continued to express MCPyV LT-Ag upon sub-
culturing, although at lower levels than those for SV40 LT-Ag-
transformed cells (Fig. 12B). However, in contrast to the trun-
cated protein, isolated pBRK colonies expressing the full-length
MCPyV LT protein could not be passaged more than 3 or 4 times.
As shown by the example of MCCL-3-derived LT-Ags in Fig. 12C,
we also observed that cells transformed by MCPyV T-Ags are ca-
pable of anchorage-independent growth in soft agar, yet they gen-
erally form smaller colonies and grow significantly slower than
their SV40-transformed counterparts. These results were con-
firmed by an anchorage-independent growth assay using low-pas-
sage-number NIH 3T3 cells expressing LT proteins (Fig. 12D).
Journal of Virology
 Functional Characterization of the MCPyV LT Antigen

Downloaded from http://jvi.asm.org/ on June 17, 2019 by guest

FIG 9 Transcription of E2F-dependent genes cdc2 and cyclin A2 is upregulated in cells overexpressing truncated MCPyV tLT-Ag. PFSK-1 cells were transiently
transfected with the indicated LT expression plasmids or a control plasmid. At 48 h posttransfection, total RNA was isolated and real-time RT-PCR was
performed (A). Values were normalized against two housekeeping genes, and values for the control plasmid were set to 1. Values represent the averages from four
independent experiments, with error bars indicating the standard deviations. P values indicate a significant difference from results for the vector control. (B)
Western blot of the transiently transfected cells used in panel A. Arrows indicate full-length LT-Ag or truncated LT-Ag.

In summary, we show for the first time that tumor-derived
MCPyV T-Ag genes are oncogenes that induce substantial altera-
tions in morphology and growth behavior of primary rodent ep-
ithelial cells, despite the inability of tLT-Ags to bind to p53 or
reduce p53-dependent transcription.
DISCUSSION
In this study, we performed a biochemical, molecular, and functional
characterization of full-length and MCC-derived truncated MCPyV
LT proteins. Key findings of our study include the facts that (i) trun-
cated LT-Ags bind with high affinity to Rb and release transcriptional
repression, (ii) full-length, but not truncated, MCPyV LT-Ag binds
indirectly to p53 and interferes with p53-dependent transcriptional
activation, and (iii) full-length and MCC-derived T-Ag genes can
induce the transformation of primary rat epithelial cells.
MCPyV is the only human polyomavirus known to induce
tumors in its host. A number of findings suggest a causative role of
MCPyV in Merkel cell carcinoma tumorigenesis: the virus can be
identified in the majority of MCC biopsy specimens (3, 4, 6–9),
MCPyV DNA is monoclonally integrated in the tumor cells (3, 9),
the viral T-Ags are expressed in tumor cells (5, 10), and LT-Ag
sequences isolated from tumor cells harbor tumor-specific muta-

FIG 10 Binding affinity of truncated LT proteins to Rb pocket domain determined by microscale thermophoresis. Binding curves of bacterially expressed and
purified His-tagged MCPyV LT1-244 (A) or SV40 LT7-117 (B) binding to the Rb pocket domain Rb382-771 at a 150 mM salt concentration (conc.). Fitting curve and
parameters are depicted.

March 2014 Volume 88 Number 6 jvi.asm.org 3155

Borchert et al.

TABLE 2 Kd values determined by microscale thermophoresisa quent cell death and is qualitatively different from a knockdown of
Kd value for Rb382–771 sT-Ag only (26). Interestingly, however, Shuda and colleagues re-
cently demonstrated that approximately 5 to 10% of MCCs ex-
Construct 150 mM NaCl 250 mM NaCl
press sT antigen but are negative for LT protein expression (25).
SV40 LT7-117 1950 nM (367.5) — Since sT-Ag, but not LT-Ag, was also found to be able to transform
MCPyV LT-Ag1-244 51.3 nM (38.1) 121 nM (35.2) rat-1 cells (25), it could be argued that sT-Ag may represent the
a
Average Kd value of three independent measurements by applying microscale major oncogene of MCPyV. In such a scenario, it also appears
thermophoresis. Standard deviations are indicated in parentheses. —, no binding.
possible that preservation of the LxCxE motif in LT-Ag is an indi-
rect result of the pressure to maintain a functional sT-Ag gene.
tions that lead to premature truncation of LT-Ag, presumably due Our survey and a statistical analysis of truncated LT-Ag coding
to a selection pressure to abrogate viral replication (9, 18, 64, 65). regions, however, clearly indicate that MCC development selects
It is striking that all truncated LT proteins retain the LxCxE motif, for the presence of an amino-terminal region that significantly
a fact that would seem to provide a strong argument for a domi- extends beyond the first exon shared with the sT-Ag gene and
nant requirement for LT-Ag-mediated Rb inactivation during which includes the LxCxE motif. As this selection also applies to
MCC pathogenesis. Support for this hypothesis comes from the the integrated genomes of those MCC tumors that have ceased to

Downloaded from http://jvi.asm.org/ on June 17, 2019 by guest

observation that knockdown of pan-T-Ag expression in sT- and express LT-Ag (25), it appears likely that unequivocal expression
LT-positive MCC cell lines results in growth arrest and subse- of truncated, LxCxE-containing LT proteins is of critical impor-

FIG 11 Subcellular localization of Rb protein in cells transiently transfected with truncated MCPyV tLT-Ag. H1299 cells were transiently transfected with Rb and
the indicated LT expression plasmids. At 48 h posttransfection, cells were stained with Cm2B4 MAb, Pab419 MAb, and FITC (Rb)/TRITC (LT)-conjugated
secondary antibodies.

3156 jvi.asm.org Journal of Virology

 Functional Characterization of the MCPyV LT Antigen

Downloaded from http://jvi.asm.org/ on June 17, 2019 by guest

FIG 12 Transformation of primary baby rat kidney (pBRK) cells. (A) pBRK cells were transiently transfected with pCMV2 plasmids expressing full-length
MCPyV LTFL-Ag, truncated MCPyV tLT-Ag protein, or SV40 LTFL-Ag. Colonies were quantified 28 days posttransfection. P values using an unpaired test are
indicated. The experiment was performed six times in duplicate. (B) Western blot analysis of individual cell clones isolated in panel A. Thirty-five micrograms
of total protein extract was loaded for MCPyV LT-expressing pBRK cells, while for SV40 LT expressing cells, only 10 ␮g total protein was loaded. LT-Ag proteins
are highlighted with an arrow; the asterisk indicates the position of SV40 LT multimers. (C and D) Anchorage-independent growth of pBRK cell lines
transformed with SV40 LTFL or truncated MCPyV tLT-MCCL-3 (C) and NIH 3T3 cells stably expressing LT proteins (D); individual pictures were taken with
phase contrast (⫻20) 28 days past seeding.

tance especially during the early steps of MCC tumorigenesis. At
later stages, a subset of MCC tumors may become independent of
continuous LT-Ag expression, perhaps as a result of the acquisi-
tion of additional mutations within the cellular genome and im-
munological pressure to limit viral gene expression. Recent pub-
lications by Cheng et al. and Li et al. demonstrated by using
immortalized cells that truncated MCPyV-LT proteins, in con-
trast to full-length LT protein, efficiently promote cell growth (21,
29); this result suggests that the MCC-specific truncations of
LT-Ag may serve to abrogate growth inhibitory properties map-
ping to the C terminus of MCPyV LT protein.
In contrast to the preserved LxCxE motif, our analysis suggests
that there is no positive selection for the presence of a functional
NLS in MCC-derived tLT-Ags. At a calculated molecular mass of
30.5 kDa or less, truncated LT proteins that lack an NLS should
nevertheless be able to enter the nucleus via passive diffusion, a
hypothesis that is supported by our localization studies. Interest-
ingly, besides the nuclear LT-Ag fraction, we find that a consider-
able subfraction of not only NLS-deficient but also truncated LT-
Ags that still contain an NLS localizes to the cytoplasm. Given the
fact that similar observations were made for the MCC-derived cell
line MKL-1 (which likewise expresses a tLT that still contains the
reported NLS), one might argues against the possibility that the
partial cytoplasmic localization occurs only during ectopic ex-
pression. This observation raises the possibility that an additional
feature of truncated LT-Ags may be their ability to relocalize in-
teraction partners to the cytoplasm. We have shown through ex-
amples that this is the case for ectopically and endogenously ex-
pressed Rb protein. Of note, we think it unlikely that the partial
cytoplasmic relocalization of Rb itself has functional conse-
quences, given that full-length (and strictly nuclear) LT protein
efficiently relieves Rb-mediated transcriptional repression. It thus
may be more interesting to consider such a mechanism for other
interaction partners of truncated LT-Ags, e.g., cellular proteins

March 2014 Volume 88 Number 6 jvi.asm.org 3157

Borchert et al.
that bind to the unique ⬃110-amino-acid region which precedes
the LxCxE region. Another interesting case is that of interaction
partners that relocalize to the nucleus when complexed with full-
length LT-Ag but fail to do so upon interaction with the truncated
protein, such as the lysosomal processing regulator hVam6p (66).
It should also be noted that transformation by cytoplasmic
LT-Ags is not without precedent, given that artificially generated
mutants of SV40 LT-Ag that fail to localize to the nucleus have
been found to be capable of transforming continuously growing
rodent cells (67).
In SV40, binding to Rb and p53 contributes to LT protein
functions during viral replication and cellular transformation.
SV40 LT protein binds directly to the DNA binding domain of
p53, thereby suppressing p53-mediated transcriptional activa-
tion. The C-terminal domain of LT contains two regions, box 1
and box 2; based on the X-ray structure of SV40 LT in complex
with p53, these regions form a single p53 binding site (68). Al-
though amino acids contributing to this particular interaction are
not well conserved in MCPyV LT (Fig. 1), earlier experiments
using SV40 LT have shown that T-Ag can additionally block p53-
dependent transcription and growth arrest by mechanisms that
are independent of p53 binding; the N-terminal 121 aa are suffi-
cient to mediate this effect (69, 70). In light of these findings, we
have analyzed full-length and truncated MCPyV LT-Ags for their
ability to bind to p53 and block p53-dependent transcription. In-
terestingly, we found that p53-dependent transcription is re-
pressed by full-length MCPyV LT-Ag. Furthermore, we observed
that a minor fraction of the protein is in a complex with p53,
although significantly less protein coprecipitates p53 than that
seen with SV40 LT. Although we cannot fully exclude the possi-
bility that our MCPyV FACS-FRET results were false negative
(e.g., due to sterical or structural hindrances), the absence of a
signal in this assay together with the poor conservation of residues
mediating the contact between SV40 LT-Ag and p53 suggest that a
subfraction of MCPyV is complexed with p53 via an indirect in-
teraction. The fact that truncated LT proteins do not coprecipitate
p53 suggests that the binding site for such a partner is likely lo-
cated in a carboxy-terminal region of LT-Ag.
We did not observe an increase in p53 steady-state levels when
MCPyV LT full-length protein was expressed, as was recently re-
ported by Li and colleagues (29). Likewise, we were unable to
detect increased p21 protein expression or an elevated level of p21,
GADD45A, or HDM2 transcription in PFSK-1 or U2OS cells
(data not shown). Neither Li et al. nor Cheng et al. (21) observed
an interaction between full-length MCPyV LT and p53. At pres-
ent, we do not have an explanation for these discrepant observa-
tions other than the use of different cell lines or experimental
approaches. Clearly, however, the results of the three studies are in
agreement that shortened MCPyV LT proteins are unable to bind
to p53 and interfere with its functions.
Despite their inability to bind to p53, we show here for the first
time that the MCC-derived early region expressing a shortened LT
can transform primary rodent cells. Similar to data obtained for
the BK and JC virus early regions (15–17), however, the efficiency
with which these cells were transformed was substantially lower
than that for SV40. Likewise, although isolated pBRK cell clones
transformed with truncated MCPyV LT proteins do express LT,
they do so at significantly lower levels than pBRK cells trans-
formed with SV40 LT. In addition, these cells proliferate much
slower than SV40 LT-transformed cells and show a relatively high
3158 jvi.asm.org
percentage of apoptotic cells (data not shown), characteristics
which have been described for MCC tumor cells (71). It should be
pointed out our experiments employ genomic clones that harbor
the entire early region from wild-type or MCC-derived MCPyV
genomes. These constructs thus should be able to express not only
LT but also sT, and it is likely that the sT protein contributes to the
transformation events observed in our assays. We have not explic-
itly tested the transforming potential of sT-deficient LT expres-
sion constructs (such constructs would also be expected to yield
significantly lower LT levels due to the absence of sT-mediated
stabilization of the LT protein). However, given that only con-
structs encoding truncated LT proteins yield continuously grow-
ing colonies in focus formation assays, and since tLT expression
furthermore resulted in the outgrowth of larger colonies in an-
chorage-independent growth assays, our data suggest an impor-
Downloaded from http://jvi.asm.org/ on June 17, 2019 by guest
tant role especially of truncated LT protein during the transfor-
mation of primary rodent cells. In contrast to the indirect nature
of the p53 interaction, our MST analysis shows that the interaction
of MCPyV with Rb is direct and highly efficient. Interestingly, the
dissociation constant (Kd) with which a truncated MCPyV LT-Ag
bound to the Rb pocket domain indicates an approximately 20-
fold-higher binding affinity than that of an N-terminal SV40 frag-
ment (it should be noted that the significance of this observation is
presently unclear, given that the SV40 fragment is not a naturally
occurring isolate and the full-length protein may exhibit an al-
tered binding affinity). Unfortunately, full-length MCPyV LT-Ag
was not efficiently expressed in bacteria, and we were thus unable
to directly compare its binding affinity to that of the shortened
MCC-derived proteins. Our coimmunoprecipitation experi-
ments together with real-time PCR experiments monitoring E2F-
dependent transcription of cdc2 and cyclinA2, however, suggest
that the truncated MCC-derived LT-Ag proteins indeed bind
more efficiently than the full-length protein. If so, it seems possi-
ble that the higher binding affinity may partially compensate for
the loss of p53 binding during the transformation process. Clearly,
however, under our experimental conditions, under which trun-
cated and full-length LT proteins are ectopically expressed at sim-
ilar levels, the efficiency with which the shortened proteins led to
the derepression of an E2F-regulated reporter was only slightly
higher than that for the full-length protein (Fig. 8). Nevertheless,
we suspect that, especially for MCC tissues that express LT-Ag at
very low or even undetectable levels (25), even small differences in
binding affinity may have substantial consequences during the
transformation process in vivo. Thus, it is conceivable that inhib-
itors targeting this interaction might be an effective new approach
for MCC therapy.
ACKNOWLEDGMENTS
This work was supported by the Structure and Dynamics in Infection
graduate school funded by the Federal State Hamburg, Germany.
We are very grateful to Pat Moore and Yuan Chang, University of
Pittsburgh, USA, for sending us Cm2B4 antibody. We thank Roland Hou-
ben, University Hospital of Würzburg, Germany, for sending us WaGa
cells; we thank Franz Oswald, University Hospital Ulm, Germany, for
providing the Rb and E2F expression constructs and the pI-H2A-68 lucif-
erase reporter plasmids. Furthermore, we thank Wolfgang Deppert for
providing the SV40 plasmid pZIPTEX and the antibody Pab419. We are
grateful to Wilhelm Ching for generously providing the p53-CFP/YFP
expression constructs. We thank Bernd Zobiak, HEXT, UKE Microscopic
Imaging Facility (umif) for help with confocal microscopy.
Journal of Virology

REFERENCES
1. DeCaprio JA, Garcea RL. 2013. A cornucopia of human polyomaviruses.
Nat. Rev. Microbiol. 11:264 –276. http://dx.doi.org/10.1038/nrmicro2992.
2. Korup S, Rietscher J, Calvignac-Spencer S, Trusch F, Hofmann J,
Moens U, Sauer I, Voigt S, Schmuck R, Ehlers B. 2013. Identification of
a novel human polyomavirus in organs of the gastrointestinal tract. PLoS
One 8:e58021. http://dx.doi.org/10.1371/journal.pone.0058021.
3. Feng H, Shuda M, Chang Y, Moore PS. 2008. Clonal integration of a
polyomavirus in human Merkel cell carcinoma. Science 319:1096 –1100.
http://dx.doi.org/10.1126/science.1152586.
4. Becker JC, Houben R, Ugurel S, Trefzer U, Pfohler C, Schrama D. 2009.
MC polyomavirus is frequently present in Merkel cell carcinoma of Euro-
pean patients. J. Invest. Dermatol. 129:248 –250. http://dx.doi.org/10
.1038/jid.2008.198.
5. Busam KJ, Jungbluth AA, Rekthman N, Coit D, Pulitzer M, Bini J,
Arora R, Hanson NC, Tassello JA, Frosina D, Moore P, Chang Y. 2009.
Merkel cell polyomavirus expression in Merkel cell carcinomas and its
absence in combined tumors and pulmonary neuroendocrine carcino-
mas. Am. J. Surg. Pathol. 33:1378 –1385. http://dx.doi.org/10.1097/PAS
.0b013e3181aa30a5.
6. Foulongne V, Kluger N, Dereure O, Brieu N, Guillot B, Segondy M. 2008.
Merkel cell polyomavirus and Merkel cell carcinoma, France. Emerg. Infect.
Dis. 14:1491–1493. http://dx.doi.org/10.3201/eid1409.080651.
7. Garneski KM, Warcola AH, Feng Q, Kiviat NB, Leonard JH, Nghiem P.
2009. Merkel cell polyomavirus is more frequently present in North
American than Australian Merkel cell carcinoma tumors. J. Invest. Der-
matol. 129:246 –248. http://dx.doi.org/10.1038/jid.2008.229.
8. Kassem A, Schopflin A, Diaz C, Weyers W, Stickeler E, Werner M, Zur
Hausen A. 2008. Frequent detection of Merkel cell polyomavirus in hu-
man Merkel cell carcinomas and identification of a unique deletion in the
VP1 gene. Cancer Res. 68:5009 –5013. http://dx.doi.org/10.1158/0008
-5472.CAN-08-0949.
9. Shuda M, Feng H, Kwun HJ, Rosen ST, Gjoerup O, Moore PS, Chang
Y. 2008. T antigen mutations are a human tumor-specific signature for
Merkel cell polyomavirus. Proc. Natl. Acad. Sci. U. S. A. 105:16272–16277.
http://dx.doi.org/10.1073/pnas.0806526105.
10. Shuda M, Arora R, Kwun HJ, Feng H, Sarid R, Fernandez-Figueras MT,
Tolstov Y, Gjoerup O, Mansukhani MM, Swerdlow SH, Chaudhary
PM, Kirkwood JM, Nalesnik MA, Kant JA, Weiss LM, Moore PS, Chang
Y. 2009. Human Merkel cell polyomavirus infection I. MCV T antigen
expression in Merkel cell carcinoma, lymphoid tissues and lymphoid tu-
mors. Int. J. Cancer 125:1243–1249. http://dx.doi.org/10.1002/ijc.24510.
11. Cherington V, Morgan B, Spiegelman BM, Roberts TM. 1986. Recom-
binant retroviruses that transduce individual polyoma tumor antigens:
effects on growth and differentiation. Proc. Natl. Acad. Sci. U. S. A. 83:
4307– 4311. http://dx.doi.org/10.1073/pnas.83.12.4307.
12. Kornbluth S, Cross FR, Harbison M, Hanafusa H. 1986. Transforma-
tion of chicken embryo fibroblasts and tumor induction by the middle T
antigen of polyomavirus carried in an avian retroviral vector. Mol. Cell.
Biol. 6:1545–1551.
13. Pipas JM. 2009. SV40: cell transformation and tumorigenesis. Virology
384:294 –303. http://dx.doi.org/10.1016/j.virol.2008.11.024.
14. Sullivan CS, Pipas JM. 2002. T antigens of simian virus 40: molecular
chaperones for viral replication and tumorigenesis. Microbiol. Mol. Biol.
Rev. 66:179 –202. http://dx.doi.org/10.1128/MMBR.66.2.179-202.2002.
15. Bollag B, Chuke WF, Frisque RJ. 1989. Hybrid genomes of the poly-
omaviruses JC virus, BK virus, and simian virus 40: identification of se-
quences important for efficient transformation. J. Virol. 63:863– 872.
16. Bollag B, Prins C, Snyder EL, Frisque RJ. 2000. Purified JC virus T and
T= proteins differentially interact with the retinoblastoma family of tumor
suppressor proteins. Virology 274:165–178. http://dx.doi.org/10.1006
/viro.2000.0451.
17. Dyson N, Bernards R, Friend SH, Gooding LR, Hassell JA, Major EO,
Pipas JM, Vandyke T, Harlow E. 1990. Large T antigens of many poly-
omaviruses are able to form complexes with the retinoblastoma protein. J.
Virol. 64:1353–1356.
18. Fischer N, Brandner J, Fuchs F, Moll I, Grundhoff A. 2010. Detection of
Merkel cell polyomavirus (MCPyV) in Merkel cell carcinoma cell lines:
cell morphology and growth phenotype do not reflect presence of the
virus. Int. J. Cancer 126:2133–2142. http://dx.doi.org/10.1002/ijc.24877.
19. Pantulu ND, Pallasch CP, Kurz AK, Kassem A, Frenzel L, Sodenkamp
S, Kvasnicka HM, Wendtner CM, Zur Hausen A. 2010. Detection of a
March 2014 Volume 88 Number 6
Functional Characterization of the MCPyV LT Antigen
novel truncating Merkel cell polyomavirus large T antigen deletion in
chronic lymphocytic leukemia cells. Blood 116:5280 –5284. http://dx.doi
.org/10.1182/blood-2010-02-269829.
20. Arora R, Chang Y, Moore PS. 2012. MCV and Merkel cell carcinoma: a
molecular success story. Curr. Opin. Virol. 2:489 – 498. http://dx.doi.org
/10.1016/j.coviro.2012.05.007.
21. Cheng J, Rozenblatt-Rosen O, Paulson KG, Nghiem P, DeCaprio JA.
2013. Merkel cell polyomavirus large T antigen has growth-promoting
and inhibitory activities. J. Virol. 87:6118 – 6126. http://dx.doi.org/10
.1128/JVI.00385-13.
22. Asselin C, Gelinas C, Bastin M. 1983. Role of the three polyoma virus
early proteins in tumorigenesis. Mol. Cell. Biol. 3:1451–1459.
23. Asselin C, Vass-Marengo J, Bastin M. 1986. Mutation in the polyoma-
virus genome that activates the properties of large T associated with neo-
plastic transformation. J. Virol. 57:165–172.
24. Noda T, Satake M, Yamaguchi Y, Ito Y. 1987. Cooperation of middle
and small T antigens of polyomavirus in transformation of established
fibroblast and epithelial-like cell lines. J. Virol. 61:2253–2263.
Downloaded from http://jvi.asm.org/ on June 17, 2019 by guest
25. Shuda M, Kwun HJ, Feng H, Chang Y, Moore PS. 2011. Human Merkel
cell polyomavirus small T antigen is an oncoprotein targeting the 4E-BP1
translation regulator. J. Clin. Invest. 121:3623–3634. http://dx.doi.org/10
.1172/JCI46323.
26. Houben R, Shuda M, Weinkam R, Schrama D, Feng H, Chang Y,
Moore PS, Becker JC. 2010. Merkel cell polyomavirus-infected Merkel
cell carcinoma cells require expression of viral T antigens. J. Virol. 84:
7064 –7072. http://dx.doi.org/10.1128/JVI.02400-09.
27. Kwun HJ, Shuda M, Feng H, Camacho CJ, Moore PS, Chang Y. 2013.
Merkel cell polyomavirus small T antigen controls viral replication and
oncoprotein expression by targeting the cellular ubiquitin ligase
SCF(Fbw7). Cell Host Microbe 14:125–135. http://dx.doi.org/10.1016/j
.chom.2013.06.008.
28. Neumann F, Borchert S, Schmidt C, Reimer R, Hohenberg H, Fischer
N, Grundhoff A. 2011. Replication, gene expression and particle produc-
tion by a consensus Merkel cell polyomavirus (MCPyV) genome. PLoS
One 6:e29112. http://dx.doi.org/10.1371/journal.pone.0029112.
29. Li J, Wang X, Diaz J, Tsang SH, Buck CB, You J. 2013. Merkel cell
polyomavirus large T antigen disrupts host genomic integrity and inhibits
cellular proliferation. J. Virol. 87:9173–9188. http://dx.doi.org/10.1128
/JVI.01216-13.
30. Graham FL, Smiley J, Russell WC, Nairn R. 1977. Characteristics of a
human cell line transformed by DNA from human adenovirus type 5. J.
Gen. Virol. 36:59 –74. http://dx.doi.org/10.1099/0022-1317-36-1-59.
31. Mitsudomi T, Steinberg SM, Nau MM, Carbone D, D’Amico D, Bodner
S, Oie HK, Linnoila RI, Mulshine JL, Minna JD, et al. 1992. p53 gene
mutations in non-small-cell lung cancer cell lines and their correlation
with the presence of ras mutations and clinical features. Oncogene 7:171–
180.
32. Fogh J, Fogh JM, Orfeo T. 1977. One hundred and twenty-seven cultured
human tumor cell lines producing tumors in nude mice. J. Natl. Cancer
Inst. 59:221–226.
33. Brown TC, Cerutti PA. 1986. Ultraviolet radiation inactivates SV40 by
disrupting at least four genetic functions. EMBO J. 5:197–203.
34. Stieler K, Schulz C, Lavanya M, Aepfelbacher M, Stocking C, Fischer N.
2010. Host range and cellular tropism of the human exogenous gamma-
retrovirus XMRV. Virology 399:23–30. http://dx.doi.org/10.1016/j.virol
.2009.12.028.
35. Oswald F, Dobner T, Lipp M. 1996. The E2F transcription factor acti-
vates a replication-dependent human H2A gene in early S phase of the cell
cycle. Mol. Cell. Biol. 16:1889 –1895.
36. Nevels M, Tauber B, Spruss T, Wolf H, Dobner T. 2001. “Hit-and-run”
transformation by adenovirus oncogenes. J. Virol. 75:3089 –3094. http:
//dx.doi.org/10.1128/JVI.75.7.3089-3094.2001.
37. Härtl B, Zeller T, Blanchette P, Kremmer E, Dobner T. 2008. Adeno-
virus type 5 early region 1B 55-kDa oncoprotein can promote cell trans-
formation by a mechanism independent from blocking p53-activated
transcription. Oncogene 27:3673–3684. http://dx.doi.org/10.1038/sj.onc
.1211039.
38. Wimmer P, Tauber B, Spruss T, Dobner T. 2010. Adenovirus type 5
early encoded proteins of the E1 and E4 regions induce oncogenic trans-
formation of primary rabbit cells. J. Gen. Virol. 91:1828 –1833. http://dx
.doi.org/10.1099/vir.0.020537-0.
39. Wang X, Li J, Schowalter RM, Jiao J, Buck CB, You J. 2012. Bromodo-
main protein Brd4 plays a key role in Merkel cell polyomavirus DNA
jvi.asm.org 3159
Borchert et al.
replication. PLoS Pathog. 8:e1003021. http://dx.doi.org/10.1371/journal
.ppat.1003021.
40. Schreiner S, Wimmer P, Groitl P, Chen SY, Blanchette P, Branton PE,
Dobner T. 2011. Adenovirus type 5 early region 1B 55K oncoprotein-
dependent degradation of cellular factor Daxx is required for efficient
transformation of primary rodent cells. J. Virol. 85:8752– 8765. http://dx
.doi.org/10.1128/JVI.00440-11.
41. Harlow E, Crawford LV, Pim DC, Williamson NM. 1981. Monoclonal
antibodies specific for simian virus 40 tumor antigens. J. Virol. 39:861–
869.
42. Nevels M, Rubenwolf S, Spruss T, Wolf H, Dobner T. 1997. The
adenovirus E4orf6 protein can promote E1A/E1B-induced focus forma-
tion by interfering with p53 tumor suppressor function. Proc. Natl. Acad.
Sci. U. S. A. 94:1206 –1211. http://dx.doi.org/10.1073/pnas.94.4.1206.
43. Hermannstädter A, Ziegler C, Kuhl M, Deppert W, Tolstonog GV.
2009. Wild-type p53 enhances efficiency of simian virus 40 large-T-
antigen-induced cellular transformation. J. Virol. 83:10106 –10118. http:
//dx.doi.org/10.1128/JVI.00174-09.
44. Banning C, Votteler J, Hoffmann D, Koppensteiner H, Warmer M,
Reimer R, Kirchhoff F, Schubert U, Hauber J, Schindler M. 2010. A flow
cytometry-based FRET assay to identify and analyse protein-protein in-
teractions in living cells. PLoS One 5:e9344. http://dx.doi.org/10.1371
/journal.pone.0009344.
45. Eward KL, Van Ert MN, Thornton M, Helmstetter CE. 2004. Cyclin
mRNA stability does not vary during the cell cycle. Cell Cycle 3:1057–
1061. http://dx.doi.org/10.4161/cc.3.8.987.
46. Stieler K, Fischer N. 2010. Apobec 3G efficiently reduces infectivity of the
human exogenous gammaretrovirus XMRV. PLoS One 5:e11738. http:
//dx.doi.org/10.1371/journal.pone.0011738.
47. Leppard KN, Shenk T. 1989. The adenovirus E1B 55 kd protein influ-
ences mRNA transport via an intranuclear effect on RNA metabolism.
EMBO J. 8:2329 –2336.
48. Liu X, Marmorstein R. 2007. Structure of the retinoblastoma protein
bound to adenovirus E1A reveals the molecular basis for viral oncoprotein
inactivation of a tumor suppressor. Genes Dev. 21:2711–2716. http://dx
.doi.org/10.1101/gad.1590607.
49. Jerabek-Willemsen M, Wienken CJ, Braun D, Baaske P, Duhr S. 2011.
Molecular interaction studies using microscale thermophoresis. Assay
Drug Dev. Technol. 9:342–353. http://dx.doi.org/10.1089/adt.2011.0380.
50. Nakamura T, Sato Y, Watanabe D, Ito H, Shimonohara N, Tsuji T,
Nakajima N, Suzuki Y, Matsuo K, Nakagawa H, Sata T, Katano H. 2010.
Nuclear localization of Merkel cell polyomavirus large T antigen in Merkel
cell carcinoma. Virology 398:273–279. http://dx.doi.org/10.1016/j.virol
.2009.12.024.
51. Martin ME, Berk AJ. 1998. Adenovirus E1B 55K represses p53 activation
in vitro. J. Virol. 72:3146 –3154.
52. Yew PR, Liu X, Berk AJ. 1994. Adenovirus E1B oncoprotein tethers a
transcriptional repression domain to p53. Genes Dev. 8:190 –202. http:
//dx.doi.org/10.1101/gad.8.2.190.
53. Bargonetti J, Reynisdottir I, Friedman PN, Prives C. 1992. Site-specific
binding of wild-type p53 to cellular DNA is inhibited by SV40 T antigen
and mutant p53. Genes Dev. 6:1886 –1898. http://dx.doi.org/10.1101/gad
.6.10.1886.
54. Harris KF, Christensen JB, Imperiale MJ. 1996. BK virus large T antigen:
interactions with the retinoblastoma family of tumor suppressor proteins
and effects on cellular growth control. J. Virol. 70:2378 –2386.
55. Jiang D, Srinivasan A, Lozano G, Robbins PD. 1993. SV40 T antigen
abrogates p53-mediated transcriptional activity. Oncogene 8:2805–2812.
56. Peden KW, Srinivasan A, Farber JM, Pipas JM. 1989. Mutants with
changes within or near a hydrophobic region of simian virus 40 large
3160 jvi.asm.org
tumor antigen are defective for binding cellular protein p53. Virology
168:13–21. http://dx.doi.org/10.1016/0042-6822(89)90398-X.
57. Peden KW, Srinivasan A, Vartikar JV, Pipas JM. 1998. Effects of muta-
tions within the SV40 large T antigen ATPase/p53 binding domain on viral
replication and transformation. Virus Genes 16:153–165. http://dx.doi
.org/10.1023/A:1007941622680.
58. Staib C, Pesch J, Gerwig R, Gerber JK, Brehm U, Stangl A, Grummt F.
1996. p53 inhibits JC virus DNA replication in vivo and interacts with JC
virus large T-antigen. Virology 219:237–246. http://dx.doi.org/10.1006
/viro.1996.0241.
59. Dilworth SM. 1990. Cell alterations induced by the large T-antigens of
SV40 and polyoma virus. Semin. Cancer Biol. 1:407– 414.
60. Imperiale MJ, Major EO. 2007. Polyomaviruses, p 2263–2298. In Knipe
DM, Howley PM, Griffin DE, Lamb RA, Martin MA, Roizman B, Straus
SE (ed), Fields virology, 5th ed. Lippincott Williams & Wilkins, Philadel-
phia, PA.
61. Wienken CJ, Baaske P, Rothbauer U, Braun D, Duhr S. 2010. Protein-
binding assays in biological liquids using microscale thermophoresis. Nat.
Downloaded from http://jvi.asm.org/ on June 17, 2019 by guest
Commun. 1:100. http://dx.doi.org/10.1038/ncomms1093.
62. Kim HY, Ahn BY, Cho Y. 2001. Structural basis for the inactivation of
retinoblastoma tumor suppressor by SV40 large T antigen. EMBO J. 20:
295–304. http://dx.doi.org/10.1093/emboj/20.1.295.
63. Tognon M, Corallini A, Martini F, Negrini M, Barbanti-Brodano G. 2003.
Oncogenic transformation by BK virus and association with human tumors.
Oncogene 22:5192–5200. http://dx.doi.org/10.1038/sj.onc.1206550.
64. Laude HC, Jonchere B, Maubec E, Carlotti A, Marinho E, Couturaud B,
Peter M, Sastre-Garau X, Avril MF, Dupin N, Rozenberg F. 2010. Distinct
Merkel cell polyomavirus molecular features in tumour and non tumour
specimens from patients with Merkel cell carcinoma. PLoS Pathog.
6:e1001076. http://dx.doi.org/10.1371/journal.ppat.1001076.
65. Sastre-Garau X, Peter M, Avril MF, Laude H, Couturier J, Rozenberg F,
Almeida A, Boitier F, Carlotti A, Couturaud B, Dupin N. 2009. Merkel
cell carcinoma of the skin: pathological and molecular evidence for a caus-
ative role of MCV in oncogenesis. J. Pathol. 218:48 –56. http://dx.doi.org
/10.1002/path.2532.
66. Liu X, Hein J, Richardson SC, Basse PH, Toptan T, Moore PS, Gjoerup
OV, Chang Y. 2011. Merkel cell polyomavirus large T antigen disrupts
lysosome clustering by translocating human Vam6p from the cytoplasm
to the nucleus. J. Biol. Chem. 286:17079 –17090. http://dx.doi.org/10.1074
/jbc.M110.192856.
67. Zhu J, Rice PW, Gorsch L, Abate M, Cole CN. 1992. Transformation of
a continuous rat embryo fibroblast cell line requires three separate do-
mains of simian virus 40 large T antigen. J. Virol. 66:2780 –2791.
68. Ahuja D, Saenz-Robles MT, Pipas JM. 2005. SV40 large T antigen targets
multiple cellular pathways to elicit cellular transformation. Oncogene 24:
7729 –7745. http://dx.doi.org/10.1038/sj.onc.1209046.
69. Quartin RS, Cole CN, Pipas JM, Levine AJ. 1994. The amino-terminal
functions of the simian virus 40 large T antigen are required to overcome
wild-type p53-mediated growth arrest of cells. J. Virol. 68:1334 –1341.
70. Rushton JJ, Jiang D, Srinivasan A, Pipas JM, Robbins PD. 1997. Simian
virus 40 T antigen can regulate p53-mediated transcription independent
of binding p53. J. Virol. 71:5620 –5623.
71. Pectasides D, Pectasides M, Economopoulos T. 2006. Merkel cell cancer
of the skin. Ann. Oncol. 17:1489 –1495. http://dx.doi.org/10.1093/annonc
/mdl050.
72. Lilyestrom W, Klein MG, Zhang R, Joachimiak A, Chen XS. 2006.
Crystal structure of SV40 large T-antigen bound to p53: interplay between
a viral oncoprotein and a cellular tumor suppressor. Genes Dev. 20:2373–
2382. http://dx.doi.org/10.1101/gad.1456306.
Journal of Virology